Our experience over the last six years in the use of a Continuous Monitoring System (CGMS) in humans has identified a number of key issues which need to be resolved for the approach to be completely reliable. Two observation in particular warrant careful examination: 1) There can be a lag between the plasma and tissue glucose concentrations under certain conditions, and this can reduce the degree of correlation between these two sampling domains and 2) Interactions within the tissue and at the sensor/tissue interface may perturb sensor output, thus leading to misleading measurements. Thus it will be important to systematically characterize these variations under a variety of conditions typically faced by the diabetic patient. Glucose sensor function will be evaluated by developing sensors with multiple sensing elements capable of independently detecting glucose in different micro-anatomic locations and/or alternatively monitor oxygen and/or peroxide, key reactants and products, respectively, or the enzymatic reaction upon which sensor response is based. Once developed, these sensors will be employed in monocyte/microphage cell cultures and in the Anderson cage method for testing of biocompatibility. These studies will be used to optimize the sensor outer polymeric membrane composition. Sensor performance and interaction with the surrounding microenvironment will be evaluated over the short term in rats and over the long term in rabbits. Also important will be detailed studies of glucose kinetics designed to verify our "push-pull" model. The question of insulin resistance as it affects the lag between plasma and tissue glucose will be examined in rats and in type 2 diabetic subjects. Glucose clamp studies will be employed to address the role of insulin in mediating glucose transport. Finally, improvements in the sensor and in the hardware and software will be incorporated into a miniaturized glucose monitoring system that will be tested in clinical studies in Year 3. These studies will include detection of nocturnal hypoglycemia and general evaluation of system performance.